Re: [PATCH v3 2/3] spis: mediatek: add spi slave for Mediatek MT2712

From: Mark Brown
Date: Tue Sep 18 2018 - 12:31:06 EST


On Mon, Sep 17, 2018 at 10:19:21AM +0800, Leilk Liu wrote:

This looks overall pretty good, a few smallish comments below:

Please use subject lines matching the style for the subsystem. This
makes it easier for people to identify relevant patches.

> if SPI_SLAVE
> +config SPI_SLAVE_MT27XX
> + tristate "MediaTek SPI slave device"
> + depends on ARCH_MEDIATEK || COMPILE_TEST
> + help
> + This selects the MediaTek(R) SPI slave device driver.
> + If you want to use MediaTek(R) SPI slave interface,
> + say Y or M here.If you are not sure, say N.
> + SPI slave drivers for Mediatek MT27XX series ARM SoCs.
>
> config SPI_SLAVE_TIME

This is a driver not a slave implementation, it should be with the other
drivers in the Kconfig. The slave implementations are for the
functionality that uses the drivers, not the drivers themselves.

> + /* set the tx/rx endian */
> +#ifdef __LITTLE_ENDIAN
> + reg_val &= ~SPIS_TX_ENDIAN;
> + reg_val &= ~SPIS_RX_ENDIAN;
> +#else
> + reg_val |= SPIS_TX_ENDIAN;
> + reg_val |= SPIS_RX_ENDIAN;
> +#endif
> + writel(reg_val, mdata->base + SPIS_CFG_REG);

This seems weird - it looks like it's changing the endianess of the
protocol based on the endianness the processor is running in. What's
going on here? I'd expect the driver to be just treating everything as
a byte stream and letting the protocol driver handle the endianness
issues, otherwise we might end up with double converions.

> + xfer->tx_dma = dma_map_single(dev, (void *)xfer->tx_buf,
> + xfer->len, DMA_TO_DEVICE);

Why is there a cast to void here? That's usually a sign that there's a
type safety issue, the whole point with void is that it's compatible
with any other pointer.

> +static irqreturn_t mtk_spi_slave_interrupt(int irq, void *dev_id)
> +{
> + struct spi_controller *ctlr = dev_id;
> + struct mtk_spi_slave *mdata = spi_controller_get_devdata(ctlr);
> + struct spi_transfer *trans = mdata->cur_transfer;
> + u32 int_status, reg_val, cnt, remainder;
> +
> + int_status = readl(mdata->base + SPIS_IRQ_ST_REG);
> + writel(int_status, mdata->base + SPIS_IRQ_CLR_REG);
> +
> + if (!trans)
> + return IRQ_HANDLED;

Are you sure that this is the right thing to do if we get a spurious
interrupt - the normal thing would be to return IRQ_NONE, possibly
logging a warning as well?

> + if (int_status & CMD_INVALID_ST)
> + dev_err(&ctlr->dev, "cmd invalid\n");

If there's an interrupt that doesn't have any of the interrupt status
flags set I'd expect to see a warning and probably IRQ_NONE being
returned. That way if the interrupt line is shared we work correctly
and if something goes wrong and the interrupt gets stuck on then the
core interrupt code's error handling can kick in.

Attachment: signature.asc
Description: PGP signature